1 Goal of the EDA

The goal of this EDA is to understand different parameters which impacts an individual’s choice of Airpot or Airline.

2 Prerequisites for analysis

Following needs to be prior to any analysis:

Clear your environment
Load required packages
Organize work and point to the right directories

Load two dataset provided with this project.

Survey data is the record of every individual
Airport Airline data is raw data for trating missing values in survey. There are 3 sheets in this file

dat1 <- read_excel("../raw_data/airport_choice_survey_EN_ver2.0_Capstone.xlsx", sheet = 1)
dat2 <- read_excel("../raw_data/Airport_Airline_data.xlsx", sheet = 1)
airline_airfare <- read_excel("../raw_data/Airport_Airline_data.xlsx", sheet = 2)
provincial_distances <- read_excel("../raw_data/Airport_Airline_data.xlsx", sheet = 3)

3 Descriptive Statistics

Summarization of the given data

# Examine data structure 
str(dat1)

## Classes 'tbl_df', 'tbl' and 'data.frame':    488 obs. of  28 variables:
##  $ ID                       : num  1 2 3 4 5 6 7 8 9 10 ...
##  $ Airport                  : num  1 1 1 1 1 1 1 1 1 1 ...
##  $ Airline                  : num  1 1 1 1 1 1 1 1 3 1 ...
##  $ Age                      : num  49 49 25 29 34 43 43 52 24 52 ...
##  $ Gender                   : num  1 2 1 1 2 1 2 1 2 2 ...
##  $ Nationality              : num  1 1 1 1 1 1 1 1 1 1 ...
##  $ TripPurpose              : num  2 1 1 2 2 4 4 1 1 1 ...
##  $ TripDuration             : num  7 4 10 7 4 40 40 7 4 5 ...
##  $ FlyingCompanion          : num  0 4 2 2 0 1 1 1 1 1 ...
##  $ ProvinceResidence        : num  3 3 3 3 3 3 3 3 3 5 ...
##  $ GroupTravel              : num  2 2 2 2 2 2 2 2 2 1 ...
##  $ NoTripsLastYear          : num  3 1 1 2 2 4 0 0 0 2 ...
##  $ FrequentFlightDestination: chr  "1" "1" "2" "5" ...
##  $ Destination              : num  3 2 1 2 1 1 1 2 3 3 ...
##  $ FlightNo                 : chr  "KE627" "KE703" "KE897" "KE725" ...
##  $ DepartureHr              : chr  "12" "10" "11" "15" ...
##  $ DepartureMn              : num  5 10 30 20 0 45 45 15 35 0 ...
##  $ DepartureTime            : num  2 1 1 2 2 2 2 1 3 3 ...
##  $ SeatClass                : num  1 1 1 1 1 1 1 1 1 1 ...
##  $ Airfare                  : num  80 41 NA 40 40 34 34 56 34 NA ...
##  $ NoTransport              : num  1 1 2 1 1 1 1 1 2 2 ...
##  $ ModeTransport            : num  6 6 4 1 6 6 6 6 2 10 ...
##  $ AccessCost               : num  8000 8000 1000 NA 8000 8000 8000 8000 2400 90000 ...
##  $ AccessTime               : num  40 50 20 NA 50 50 50 50 5 60 ...
##  $ Occupation               : num  1 9 12 8 1 1 9 2 12 12 ...
##  $ Income                   : num  5 3 NA 7 3 3 3 3 NA NA ...
##  $ MileageAirline           : chr  "1" NA NA "2" ...
##  $ Mileage                  : num  150000 NA NA 100000 NA NA NA NA NA NA ...

summary(dat1)

##        ID           Airport        Airline           Age       
##  Min.   :  1.0   Min.   :1.00   Min.   :1.000   Min.   :17.00  
##  1st Qu.:122.8   1st Qu.:1.00   1st Qu.:1.000   1st Qu.:29.00  
##  Median :244.5   Median :2.00   Median :2.000   Median :38.00  
##  Mean   :244.5   Mean   :1.51   Mean   :2.423   Mean   :39.97  
##  3rd Qu.:366.2   3rd Qu.:2.00   3rd Qu.:4.000   3rd Qu.:50.00  
##  Max.   :488.0   Max.   :2.00   Max.   :4.000   Max.   :80.00  
##                                 NA's   :10      NA's   :1      
##      Gender       Nationality     TripPurpose     TripDuration   
##  Min.   :1.000   Min.   :1.000   Min.   :1.000   Min.   :  0.00  
##  1st Qu.:1.000   1st Qu.:1.000   1st Qu.:1.000   1st Qu.:  4.00  
##  Median :2.000   Median :1.000   Median :1.000   Median :  5.00  
##  Mean   :1.528   Mean   :1.484   Mean   :1.518   Mean   : 27.44  
##  3rd Qu.:2.000   3rd Qu.:1.000   3rd Qu.:2.000   3rd Qu.:  8.00  
##  Max.   :2.000   Max.   :5.000   Max.   :4.000   Max.   :730.00  
##  NA's   :3                                                       
##  FlyingCompanion ProvinceResidence  GroupTravel    NoTripsLastYear  
##  Min.   : 0.00   Min.   :1.000     Min.   :1.000   Min.   :  0.000  
##  1st Qu.: 1.00   1st Qu.:1.000     1st Qu.:2.000   1st Qu.:  1.000  
##  Median : 2.00   Median :3.000     Median :2.000   Median :  2.000  
##  Mean   : 2.82   Mean   :3.391     Mean   :1.826   Mean   :  3.262  
##  3rd Qu.: 3.00   3rd Qu.:5.000     3rd Qu.:2.000   3rd Qu.:  3.000  
##  Max.   :34.00   Max.   :8.000     Max.   :2.000   Max.   :122.000  
##                                                                     
##  FrequentFlightDestination  Destination      FlightNo        
##  Length:488                Min.   :1.000   Length:488        
##  Class :character          1st Qu.:1.000   Class :character  
##  Mode  :character          Median :2.000   Mode  :character  
##                            Mean   :2.176                     
##                            3rd Qu.:3.000                     
##                            Max.   :4.000                     
##                            NA's   :5                         
##  DepartureHr         DepartureMn    DepartureTime     SeatClass    
##  Length:488         Min.   : 0.00   Min.   :1.000   Min.   :1.000  
##  Class :character   1st Qu.:15.00   1st Qu.:2.000   1st Qu.:1.000  
##  Mode  :character   Median :30.00   Median :2.000   Median :1.000  
##                     Mean   :25.98   Mean   :2.432   Mean   :1.122  
##                     3rd Qu.:40.00   3rd Qu.:3.000   3rd Qu.:1.000  
##                     Max.   :55.00   Max.   :4.000   Max.   :3.000  
##                     NA's   :120                     NA's   :4      
##     Airfare        NoTransport    ModeTransport      AccessCost    
##  Min.   :  3.00   Min.   :1.000   Min.   : 1.000   Min.   :     0  
##  1st Qu.: 35.00   1st Qu.:1.000   1st Qu.: 2.000   1st Qu.:  2000  
##  Median : 45.00   Median :1.000   Median : 3.000   Median :  6000  
##  Mean   : 50.46   Mean   :1.334   Mean   : 3.801   Mean   : 11220  
##  3rd Qu.: 60.00   3rd Qu.:2.000   3rd Qu.: 6.000   3rd Qu.: 12000  
##  Max.   :260.00   Max.   :4.000   Max.   :11.000   Max.   :350000  
##  NA's   :155                                       NA's   :197     
##    AccessTime       Occupation         Income      MileageAirline    
##  Min.   :  4.00   Min.   : 1.000   Min.   :1.000   Length:488        
##  1st Qu.: 25.00   1st Qu.: 2.000   1st Qu.:2.000   Class :character  
##  Median : 40.00   Median : 8.000   Median :3.000   Mode  :character  
##  Mean   : 51.83   Mean   : 6.768   Mean   :3.126                     
##  3rd Qu.: 60.00   3rd Qu.:12.000   3rd Qu.:4.000                     
##  Max.   :390.00   Max.   :12.000   Max.   :7.000                     
##  NA's   :97                        NA's   :132                       
##     Mileage      
##  Min.   :     1  
##  1st Qu.:  8258  
##  Median : 27998  
##  Mean   : 56384  
##  3rd Qu.: 61808  
##  Max.   :500000  
##  NA's   :398

We have so many missing values in survey data:

Neglecting parameters where less than 10 records are missing, DepartureMinute - flight departure time has 120 records missing
Airfare - Price of the ticket has 155 records missing
Access Cost - Cost from residence to Airport has 197 records missing
Access Time - Commute time from residence to Airport has 97 records missing
Income - Earing of the survey participant has 132 records missing
Mileage - Air miles from airlines for frequent travel customer has 398 records missing

4 Data manipulation

Changing the survey data to make it easier to read and use during EDA process

4.1 Categorical Variables - Mutation

4.1.1 Airport

Modifying the Airport Name for purpose of analysis

There are no records in survey data where Airport is missing

# 1. Inchoen (ICN) Airport
# 2. Gimpo (GMP) Airport
# No missing values
dat1 <- dat1 %>%
  mutate(Airport = case_when(Airport == 1 ~ "ICN",
                            Airport == 2 ~ "GMP"))

4.1.2 Airline

Modifying Airline Choice Names

There are 10 records in survey data where Airline is missing
We will try to get this data using Flight number if provided for these missing records, else remove them for modelling

# 1. Korean Air(KE)
# 2. Asiana Airlines (OZ)
# 3. Korean LCC
# 4. Foreign Airlines

# Replace missing values
dat1 <- dat1 %>%
  mutate(Airline = ifelse(!is.na(Airline), Airline, ifelse(str_detect(FlightNo, "KE"), "Korean Air", "Unknown"))) %>%
  filter(Airline != "Unknown")

dat1 <- dat1 %>%
 mutate(Airline = case_when(Airline == 1 ~ "Korean Air",
                           Airline == 2 ~ "Asiana Airlines",
                           Airline == 3 ~ "Korean LCC",
                           Airline == 4 ~ "Foreign Airlines"))

4.1.3 Gender

Modifying Gender

There are 3 records in survey data where Gender is missing
We will consider them and Not mentioned instead of filling with mode value

dat1 <- dat1 %>%
  mutate(Gender = if_else(Gender == 1, "Male","Female","Not Mentioned"))

4.1.4 Nationality

Modifying Nationality

There are no records in survey data where Nationality is missing

# 1. Korea
# 2. China
# 3. Japan
# 4. Southeast Asia
# 5. Other
dat1 <- dat1 %>%
  mutate(Nationality = case_when(Nationality == 1 ~ "Korea",
                            Nationality == 2 ~ "China",
                            Nationality == 3 ~ "Japan",
                            Nationality == 4 ~ "Southeast Asia",
                            Nationality == 5 ~ "Other"))

4.1.5 Trip Purpose

Modifying Purpose of Trip

There are no records in survey data where Trip Purpose is missing

# 1. Leisure Travel (sightseeing, visiting family, vacation, shopping)
# 2. Business Travel 
# 3. Study
# 4. Other
dat1 <- dat1 %>%
  mutate(TripPurpose = case_when(TripPurpose == 1 ~ "Leisure",
                            TripPurpose == 2 ~ "Business",
                            TripPurpose == 3 ~ "Study",
                            TripPurpose == 4 ~ "Other"))

4.1.6 Province Residence

Modifying Province Residence

There are no records in survey data where Province Residence is missing but access time and distance from province is missing for some records

4.1.6.1 Data manipulation in Province raw data file to fill missing values in survey

provincial_distances <- provincial_distances %>% 
  separate(Province, c("ProvinceResidence", "Province"), sep = "\\.", remove=FALSE)

4.1.6.2 Add province distance and time to the airport_survey_data : Join both raw data

# Join the two datasets on airport name and province ids

dat1 <- dat1 %>%
  mutate(ProvinceResidence = as.character(ProvinceResidence)) %>%
  left_join(provincial_distances, by = c("Airport", "ProvinceResidence")) %>%
  select(-c(ProvinceResidence)) %>%
  mutate(Province = as.factor(Province),
          Airport = as.factor(Airport))

4.1.7 Group Travel

Modifying Group Travel

There are no records in survey data where Group Travel is missing

# 1. Yes
# 2. No
dat1 <- dat1 %>%
  mutate(GroupTravel = case_when(GroupTravel == 1 ~ 1, GroupTravel == 2 ~ 0))

4.1.8 FrequentFlightDestination

Modifying Frequent Travel Destinations

There are no records in survey data where Frequent Travel Destinations is missing

# 1. Southeast Asia
# 2. China
# 3. Japan
# 4. North/South America
# 5. Europe
# 6. Other
# 7. None
dat1 <- dat1 %>%
  mutate(FrequentFlightDestination = case_when(FrequentFlightDestination == 1 ~ "Southeast Asia",
                            FrequentFlightDestination == 2 ~ "China",
                            FrequentFlightDestination == 3 ~ "Japan",
                            FrequentFlightDestination == 4 ~ "North/South America",
                            FrequentFlightDestination == 5 ~ "Europe",
                            FrequentFlightDestination == 6 ~ "Other",
                            FrequentFlightDestination == 7 ~ "None"))

4.1.9 Destination

Modifying Flight Destinations

There are 5 records in survey data where Flight Destinations of individual is missing
can we get information from flight No?

# 1. China
# 2. Japan
# 3. Southeast Asia
# 4. Other
dat1 <- dat1 %>%
  mutate(Destination = case_when(Destination == 1 ~ "China",
                            Destination == 2 ~ "Japan",
                            Destination == 3 ~ "Southeast Asia",
                            Destination == 4 ~ "Other"))

4.1.10 DepartureTime

Modifying Departure Time

There are no records in survey data where Departure Times is missing

# 1. 6am - 12pm
# 2. 12pm - 6pm
# 3. 6pm - 9pm
# 4. 9pm - 6am
dat1 <- dat1 %>%
  mutate(DepartureTime = case_when(DepartureTime == 1 ~ "6am - 12pm",
                            DepartureTime == 2 ~ "12pm - 6pm",
                            DepartureTime == 3 ~ "6pm - 9pm",
                            DepartureTime == 4 ~ "9pm - 6am"))

4.1.11 SeatClass

Modifying Seat Class

There are 4 records in survey data where Seat Class is missing
If we will use this parameter in some modelling, we can fill missings with mode.

# 1. Economy
# 2. Business
# 3. First Class
dat1 <- dat1 %>%
  mutate(SeatClass = case_when(SeatClass == 1 ~ "Economy",
                            SeatClass == 2 ~ "Business",
                            SeatClass == 3 ~ "First Class"))

4.1.12 ModeTransport

Modifying Mode of Transport

There are no records in survey data where Mode of Transport is missing
We are removing Air transport mode - that is outlier for this parameter

# 1. Car
# 2. Taxi
# 3. Bus
# 4. Express Bus
# 5. Subway
# 6. Airport Limousine
# 7. Rail
# 8. KTX (Express Rail)
# 9. Shuttle provided by travel agency
# 10. Air
# 11. Other
dat1 <- dat1 %>%
  mutate(ModeTransport = case_when(ModeTransport == 1 ~ "Private Car",
                                   ModeTransport == 2 ~ "Private Taxi",
                                   ModeTransport == 3 ~ "Public Bus",
                                   ModeTransport == 4 ~ "Public Bus",
                                   ModeTransport == 5 ~ "Subway",
                                   ModeTransport == 6 ~ "Private Taxi",
                                   ModeTransport == 7 ~ "Public Train",
                                   ModeTransport == 8 ~ "Public Train",
                                   ModeTransport == 9 ~ "Private Taxi",
                                   ModeTransport == 11 ~ "Other")) %>%
  filter(ModeTransport != 10)
# Regroup into maximum 3-4 buckets

4.1.13 Age

Modifying Age and add them into groups

#Youth (15-24 years) 
#Adults (25-64 years) 
#Seniors (65 years and over)
# dat1 <- dat1 %>%
#    mutate(AgeGroup = case_when(Age >=15 & Age <= 24 ~ "15-24",
#                                Age >=25 & Age <= 34 ~ "25-34",
#                                Age >=35 & Age <= 44 ~ "35-44",
#                                Age >=45 & Age <= 54 ~ "45-54",
#                                Age >=55 & Age <= 64 ~ "55-64",
#                                TRUE ~ "Senior"))

4.1.14 Occupation

Modifying Occupation

# 1. Entrepreneur, Senior management
# 2. Business (Corporate worker)
# 3. Sales, Service
# 4. Government, Military
# 5. Professionals (doctor, lawyer, professor)
# 6. industrial, manufacturing
# 7. self-employeed
# 8. Student
# 9. Housewife
# 10. Retired
# 11. None
# 12. Other

4.1.15 Income

Modifying Income Class

# 1. 30 Million Won or less
# 2. 30~50 Million Won
# 3. 50~80 Million Won
# 4. 80 ~100 Million Won
# 5. 100 ~150 Million Won
# 6. 150 ~200 Million Won
# 7. 200 Million Won or more

4.1.16 MileageAirline

Modifying Mileage Airline

There are no records in survey data where Mileage Airline is missing

# 1. Korea Airlines
# 2. Asian Airlines
# 3. Others
# 4. None
# 5. Unknown Missing Value
dat1 <- dat1 %>%
  mutate(MileageAirline = case_when(MileageAirline == 1 ~ "Korea Airlines",
                                   MileageAirline == 2 ~ "Asian Airlines",
                                   MileageAirline == 3 ~ "Others",
                                   MileageAirline == 4 ~ "None",
                                   MileageAirline == 5 ~ "Unknown Missing Value"))
#dat1 <- dat1 %>%
#  mutate(MileageAirline = ifelse(is.na(Mileage), 'No Mileage', 'Mileage'))

4.2 Quantitative Variables - Handle Missing values

4.2.1 Age

Replacing missing values in age

There is 1 record where Age is missing in survey data provided
We have to check the distribution of Age before filling the missing value - So histogram

# Check distribution of age
hist(dat1$Age)

The distribution appears to be normal and not highly skewed/ outliers.
Hence, We use mean of the age of survey participants to replace missing entries

age_mean = mean(dat1$Age, na.rm = TRUE)
dat1$Age[is.na(dat1$Age)] <- age_mean

After this operation we are left with no records where this Age is missing

4.2.2 Trip Duration

4.2.3 Flying Companion

Not relevant

4.2.4 NoTripsLastYear

4.2.5 Airfare

Replacing missing values in Airfare

There are 155 records where Air fare is missing in survey data provided
We have to check the distribution of Airfare before filling the missing value - So histogram
We also have Airport_Airline_data -> raw data which can be used to treat missing values

# Check distribution of Airfare
hist(dat1$Airfare)

# change column name for join
names(airline_airfare)[names(airline_airfare) == 'Flight.Destination'] <- 'Destination'

dat1 = dat1 %>%
  left_join(airline_airfare, by = c("Airport", "Airline", "Destination")) %>%
  mutate(Airfare = ifelse(is.na(dat1$Airfare), `Average Price`/10000, Airfare))

# fill remaining from survey data average of group of -- Airport, Airlines, Destination, seat type
airfare_average = dat1 %>%
  group_by(Airport, Airline, Destination, SeatClass) %>%
  summarise(a_fare=mean(Airfare, na.rm = TRUE))

dat1 = dat1 %>%
  left_join(airfare_average, by = c("Airport", "Airline", "Destination", "SeatClass")) %>%
  mutate(Airfare = ifelse(is.na(dat1$Airfare), a_fare, Airfare))

Airfare in raw data file is multiplied by 10000, so to use that value we have to divide it by 10000
After this operation we are left with 6 records where this airfare is missing

4.2.6 NoTransport

4.2.7 AccessCost

Replacing missing values in Access Cost

There are 197 records where Access Cost is missing in survey data provided
We have to check the distribution of Access Cost before filling the missing value - So histogram

# Check distribution of AccessCost
hist(dat1$AccessCost)

We dont have any useful information for this parameter in Airport_Airline_data -> raw data so these can not be treated
Also, 197 is a almost 40% of total records, so we are not going to include this variable in any model

4.2.8 AccessTime

Replacing missing values in Access time

There are 97 records where Access time is missing in survey data provided
We have to check the distribution of Access time before filling the missing value - So histogram
We also have Airport_Airline_data -> raw data which can be used to treat missing values

# Check distribution of AccessTime
hist(dat1$AccessTime)

Replace missing values in Access time

dat1 = dat1 %>%
mutate(AccessTime = ifelse(is.na(dat1$AccessTime),
                             `Travel Time, minutes`,
                             AccessTime))

Access time in raw data file is given in the same units (Minutes), so we use that same value to treat missing, irrespective of no of transport modes used
After this operation we are left with no records where this Access time is missing

4.2.8.1 Mileage

Replacing missing values in Mileage

There are 398 records where Mileage is missing in survey data provided
We have to check the distribution of Mileage before filling the missing value - So histogram
We also have Airport_Airline_data -> raw data which can be used to treat missing values

# Check distribution of Mileage
hist(dat1$Mileage)

We dont have any useful information for this parameter in Airport_Airline_data -> raw data so these missings can not be treated
Also, 398 is a almost 80% of total records, so we are not going to include this variable in any model

4.2.9 Replace remaining NAs by 999 for purpose of EDA

# There are only 25 rows without NA, so instead of removing NAs replace it with 999
dat1[is.na(dat1)] <- 999

4.3 Adjust Dataset

4.3.1 Removing unnecessary columns

Remove ID column and Flight No column as they do not provide any benefit in this analysis because of their cardinal
Remove Departure Hr and Min as information is already captured in Departure Time
Remove Occupation - Not relevant for this study
Remove Mileage - 398 missing values
Remove MileageAirline - Missing values or unknown values
Remove AccessCost, Income - Lot of missing values
Remove FlyingCompanion, NoTripsLastYear - Not relevant for this study

# Remove ID column and Flight No column as they do not provide any benefit in this analysis because of their cardinal
dat1 <- dat1 %>%
  select(-c(FlightNo, ID))

# Remove Departure Hr and Min as information is already captured in Departure Time
dat1 <- dat1 %>%
  select(-c(DepartureHr, DepartureMn))

# Remove Occupation - Not relevant for this study
dat1 <- dat1 %>%
  select(-c(Occupation))

# Remove Mileage - 398 missing values
# Remove MileageAirline - Missing values or unknown values
dat1 <- dat1 %>%
  select(-c(Mileage, MileageAirline))

# Remove AccessCost - Lot of missing values
# Remove Income - missing values
dat1 <- dat1 %>%
  select(-c(Income))

# Not Relevant for our analysis
dat1 <- dat1 %>%
  select(-c(FlyingCompanion, NoTripsLastYear))

4.3.2 Getting the modified dataset as airport_survey_data

# Modify data
airport_survey_data <- dat1 %>%
  mutate(Airport = as.factor(Airport),
         Airline = as.factor(Airline),
         Nationality = as.factor(Nationality),
         TripPurpose = as.factor(TripPurpose),
         Province = as.factor(Province),
         Destination = as.factor(Destination),
         DepartureTime = as.factor(DepartureTime),
         SeatClass = as.factor(SeatClass),
         NoTransport = as.factor(NoTransport),
         ModeTransport = as.factor(ModeTransport),
         Gender = as.factor(Gender))

summary(airport_survey_data)

##  Airport               Airline         Age                  Gender   
##  GMP:245   Asiana Airlines :106   Min.   :17.00   Female       :249  
##  ICN:229   Foreign Airlines:137   1st Qu.:29.00   Male         :222  
##            Korean Air      :150   Median :37.00   Not Mentioned:  3  
##            Korean LCC      : 81   Mean   :39.83                      
##                                   3rd Qu.:50.00                      
##                                   Max.   :80.00                      
##                                                                      
##          Nationality    TripPurpose   TripDuration     GroupTravel    
##  China         : 32   Business:102   Min.   :  0.00   Min.   :0.0000  
##  Japan         : 40   Leisure :317   1st Qu.:  4.00   1st Qu.:0.0000  
##  Korea         :368   Other   : 29   Median :  5.00   Median :0.0000  
##  Other         : 14   Study   : 26   Mean   : 24.59   Mean   :0.1624  
##  Southeast Asia: 20                  3rd Qu.:  8.00   3rd Qu.:0.0000  
##                                      Max.   :730.00   Max.   :1.0000  
##                                                                       
##  FrequentFlightDestination         Destination     DepartureTime
##  Length:474                999           :  2   12pm - 6pm:205  
##  Class :character          China         :131   6am - 12pm: 46  
##  Mode  :character          Japan         :155   6pm - 9pm :194  
##                            Other         : 22   9pm - 6am : 29  
##                            Southeast Asia:164                   
##                                                                 
##                                                                 
##        SeatClass      Airfare       NoTransport      ModeTransport
##  999        :  4   Min.   :  3.00   1:339       Other       :  2  
##  Business   : 34   1st Qu.: 36.22   2:120       Private Car :106  
##  Economy    :424   Median : 46.83   3: 14       Private Taxi:176  
##  First Class: 12   Mean   : 61.20   4:  1       Public Bus  : 50  
##                    3rd Qu.: 58.00               Public Train: 39  
##                    Max.   :999.00               Subway      :101  
##                                                                   
##    AccessCost       AccessTime        Distance                 Province  
##  Min.   :     0   Min.   :  4.00   Min.   : 23.72    Seoul         :182  
##  1st Qu.:   999   1st Qu.: 30.00   1st Qu.: 46.00    Kyungki-do    :126  
##  Median :  1300   Median : 50.00   Median : 61.00    Jeju          : 89  
##  Mean   :  6785   Mean   : 62.36   Mean   :154.93    Kyungsang-do  : 27  
##  3rd Qu.:  8000   3rd Qu.: 64.00   3rd Qu.:363.00    Incheon       : 22  
##  Max.   :350000   Max.   :390.00   Max.   :455.00    Chungcheong-do: 13  
##                                                     (Other)        : 15  
##  Travel Time, minutes Average Price        a_fare      
##  Min.   : 39          Min.   :   999   Min.   : 14.00  
##  1st Qu.: 56          1st Qu.:   999   1st Qu.: 42.17  
##  Median : 64          Median :   999   Median : 46.83  
##  Mean   :103          Mean   :166373   Mean   : 61.20  
##  3rd Qu.:180          3rd Qu.:362160   3rd Qu.: 52.72  
##  Max.   :311          Max.   :529260   Max.   :999.00  
##

attach(airport_survey_data)

4.3.3 Observations

249 surveyed people travelled via GMP and 239 travelled via ICN
Maximum surveyed people travelled via Korean Air i.e. count is 153
Average age of surveyed people is 40 years
Maximum surveyed people have maximum trip purpose as Leisure
Average trip duration of surveyed people is 27.44
Maximum surveyed people travelled to Southeast Asia
Maximum surveyed people are from Seoul Province
Average travel time is 105 hours

4.3.4 Export dataset to be used for modeling in Python

# Selecting relevant columns for further developement of models in Python for analysis

# airport_choice <- airport_survey_data %>%
#    select(Airport, Airline,AccessTime, DepartureTime,
#          Age,Destination,Distance,AccessCost)
# 
# airline_choice <- airport_survey_data %>%
#   select(Airline,Airport,Airfare,DepartureTime,Destination)

# write.csv(airport_choice, "../code/airport_choice.csv")
# write.csv(airline_choice, "../code/airline_choice.csv")
write.csv(airport_survey_data, "../code/airport_survey_data.csv")

5 Basic EDA

5.1 Univariate non-graphical – Categorical

# Function to get the relevant grouped data frames
# df -> dataframe
# groupVariable -> the group by column 
# output -> the resulting data frame
getGroupedUnivariateData <- function(df, groupVariable) {
  # group_by() grouped variable
  # summarise() reduces variable to descriptive stat
  # count is the new variable name, n() is a counting function
  # percent is a new variable, sum() and nrow() are functions
  # tot_selling_price -> total selling price per grouped variable
  # order_by() tot_selling_price
  result <- df %>%             
    group_by_(groupVariable) %>%     
    summarise(count = n(),   
              percent = (sum(count) / nrow(df)) * 100,
              avg_airfare = (mean(Airfare)),
              avg_accessCost = (mean(AccessCost)),
              avg_accessTime = (mean(AccessTime))) %>%
    arrange(desc(percent)) %>%
    as.data.frame() 
  
  return(result)
}

5.1.1 Preferred Airport

# count, average access cost and average access time by airport
airport_data_airport_type <- getGroupedUnivariateData(airport_survey_data, "Airport")

airport_data_airport_type %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

Airport	count	percent	avg_airfare	avg_accessCost	avg_accessTime
GMP	245	51.68776	53.56583	6207.988	53.94694
ICN	229	48.31224	69.37374	7401.502	71.35808

Observations

Average airfare and average access cost from ICN airport is more than that of GMP airport, for surveyed people.

5.1.2 Preferred Airline

# count, average access cost and average access time by airline
airport_data_airline_type <- getGroupedUnivariateData(airport_survey_data, "Airline")

airport_data_airline_type %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

Airline	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Korean Air	150	31.64557	56.99893	8698.613	62.55333
Foreign Airlines	137	28.90295	75.49061	5087.482	56.16788
Asiana Airlines	106	22.36287	54.11425	7168.349	69.98113
Korean LCC	81	17.08861	54.09939	5608.383	62.49383

Observations

Average airfare is maximum for ‘other airlines’ (NA)

5.1.3 Gender

# count, average access cost and average access time by gender
airport_data_gender <- getGroupedUnivariateData(airport_survey_data, "Gender")

airport_data_gender %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

Gender	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Female	249	52.5316456	59.46850	5906.663	63.71084
Male	222	46.8354430	63.40278	7790.279	60.62162
Not Mentioned	3	0.6329114	42.38000	5233.333	78.66667

Observations

Average airfare is minimum for Female

5.1.4 Nationality

# count, average access cost and average access time by airport
airport_data_nationality <- getGroupedUnivariateData(airport_survey_data, "Nationality")

airport_data_nationality %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

Nationality	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Korea	368	77.637131	64.72743	6943.886	59.80707
Japan	40	8.438819	53.48309	5474.450	57.00000
China	32	6.751055	48.24279	5540.250	66.06250
Southeast Asia	20	4.219409	47.33056	9749.700	100.90000
Other	14	2.953587	40.05821	4949.357	81.21429

77% people in the survey are from Korea

5.1.5 Trip Purpose

# count, average access cost and average access time by airport
airport_data_TripPurpose <- getGroupedUnivariateData(airport_survey_data, "TripPurpose")

airport_data_TripPurpose %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

TripPurpose	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Leisure	317	66.877637	63.29966	5232.372	61.24290
Business	102	21.518987	62.16415	11337.343	65.51961
Other	29	6.118144	45.29150	9199.724	56.37931
Study	26	5.485232	49.61629	5155.308	70.23077

Approximate 66% people are traveling for leisure
More than 21% people are travelling for business purposes

5.1.6 Residence

# count, average access cost and average access time by destination
airport_data_Residence <- getGroupedUnivariateData(airport_survey_data, "Province")

airport_data_Residence %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

Province	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Seoul	182	38.396625	56.23822	4293.868	43.39011
Kyungki-do	126	26.582279	59.49765	7224.571	55.61905
Jeju	89	18.776371	81.42943	8015.191	74.59551
Kyungsang-do	27	5.696203	64.06941	21316.519	141.85185
Incheon	22	4.641350	45.64090	1919.773	33.59091
Chungcheong-do	13	2.742616	42.61244	7492.000	78.46154
Jeonra-do	10	2.109705	51.00000	3219.200	188.50000
Kangwon-do	5	1.054852	46.60000	12680.000	108.00000

More than 60% people are from Seoul and Kyungki-do province, which is near to GMP airport as compared to ICN
Approximate 18% people are from Jeju province, which is almost equi-distant from both the airports

5.1.7 Destination

# count, average access cost and average access time by destination
airport_data_destination <- getGroupedUnivariateData(airport_survey_data, "Destination")

airport_data_destination %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

Destination	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Southeast Asia	164	34.5991561	73.16723	8440.372	67.41463
Japan	155	32.7004219	51.03942	6089.342	55.20000
China	131	27.6371308	46.39340	5782.359	64.89313
Other	22	4.6413502	134.18831	4970.136	60.22727
999	2	0.4219409	35.00000	10500.000	60.00000

Southeast Asia, Japan and China are the famous destinations among survey participants

Observations

Other airlines have highest average airfare

5.1.8 Preferred Departure Time

# count, average access cost and average access time by province residence
airport_data_deptime <- getGroupedUnivariateData(airport_survey_data, "DepartureTime")

airport_data_deptime %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

DepartureTime	count	percent	avg_airfare	avg_accessCost	avg_accessTime
12pm - 6pm	205	43.248945	52.91412	5753.839	54.15122
6pm - 9pm	194	40.928270	64.80852	8519.015	72.23196
6am - 12pm	46	9.704641	88.91833	5789.913	55.00000
9pm - 6am	29	6.118144	51.71462	4046.172	66.00000

As per the survey data, most people prefer departure time during the day. 12pm – 6pm – 9pm

5.1.9 Preferred Seat Class

# count, average access cost and average access time by seat class
airport_data_seatClass <- getGroupedUnivariateData(airport_survey_data, "SeatClass")

airport_data_seatClass %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

SeatClass	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Economy	424	89.4514768	50.62049	6758.300	61.88915
Business	34	7.1729958	86.23899	7811.294	68.64706
First Class	12	2.5316456	293.16667	6650.000	66.50000
999	4	0.8438819	274.25000	1249.500	46.25000

Observations

Average airfare and access cost is highest for Business class seats
Minimum average airfare for Incheon Province

Observations

5.1.10 Preferred Mode of Transport

# count, average access cost and average access time by province residence
airport_data_modeTransport <- getGroupedUnivariateData(airport_survey_data, "ModeTransport")

airport_data_modeTransport %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

ModeTransport	count	percent	avg_airfare	avg_accessCost	avg_accessTime
Private Taxi	176	37.1308017	54.75351	10959.472	53.63068
Private Car	106	22.3628692	61.15355	2261.434	59.04717
Subway	101	21.3080169	62.00375	1521.762	51.53465
Public Bus	50	10.5485232	65.36405	2970.780	60.36000
Public Train	39	8.2278481	83.70869	17156.051	139.92308
Other	2	0.4219409	48.05000	37999.500	90.00000

Observations

Agency Shuttles provide the minimum average airfare
Car is the most preferred way of Transport Mode to the airport

5.2 Univariate graphical code – Categorical

# Code histograms using grid.arrange so can see all quant variables together
grid.arrange(
  airport_data_airport_type %>% 
    ggplot(aes(x = reorder(Airport, -count), y = count)) +  
    xlab("Airport") + 
    theme_classic() +
      geom_bar(stat = "identity", fill = "darkolivegreen3"),
  airport_data_airline_type %>% 
    ggplot(aes(x = reorder(Airline, -count), y = count)) +
    xlab("Airline") +
    theme_classic() +
      geom_bar(stat = "identity", fill = "darkolivegreen3"),
  airport_data_gender %>% 
    ggplot(aes(x = reorder(Gender, -count), y = count)) +
    xlab("Gender") +
    theme_classic() +
      geom_bar(stat = "identity", fill = "darkolivegreen3"),
  airport_data_seatClass %>% 
    ggplot(aes(x = reorder(SeatClass, -count), y = count)) +
    xlab("SeatClass") +
    theme_classic() +
      geom_bar(stat = "identity", fill = "darkolivegreen3"),
  airport_data_Residence %>% 
    ggplot(aes(x = reorder(Province, -count), y = count)) +
    xlab("Province") +
    theme_classic() +
      geom_bar(stat = "identity", fill = "darkolivegreen3"),
  airport_data_destination %>% 
    ggplot(aes(x = reorder(Destination, -count), y = count)) +
    xlab("Destination") +
    theme_classic() +
      geom_bar(stat = "identity", fill = "darkolivegreen3"),
  ncol = 2
)

Observations

GMP and ICN have almost equal number of respondents
Korean airline is the most preferred
Females are more frequent travellers
Economy class is the most preferred Seat Class
Maximum number of passengers belong to Seoul Province
Southeast Asia is the most travelled destination

5.3 Univariate non-graphical – Quantitative

5.3.1 Descriptive Statistics

# Get descriptive stats 
summary(airport_survey_data)

##  Airport               Airline         Age                  Gender   
##  GMP:245   Asiana Airlines :106   Min.   :17.00   Female       :249  
##  ICN:229   Foreign Airlines:137   1st Qu.:29.00   Male         :222  
##            Korean Air      :150   Median :37.00   Not Mentioned:  3  
##            Korean LCC      : 81   Mean   :39.83                      
##                                   3rd Qu.:50.00                      
##                                   Max.   :80.00                      
##                                                                      
##          Nationality    TripPurpose   TripDuration     GroupTravel    
##  China         : 32   Business:102   Min.   :  0.00   Min.   :0.0000  
##  Japan         : 40   Leisure :317   1st Qu.:  4.00   1st Qu.:0.0000  
##  Korea         :368   Other   : 29   Median :  5.00   Median :0.0000  
##  Other         : 14   Study   : 26   Mean   : 24.59   Mean   :0.1624  
##  Southeast Asia: 20                  3rd Qu.:  8.00   3rd Qu.:0.0000  
##                                      Max.   :730.00   Max.   :1.0000  
##                                                                       
##  FrequentFlightDestination         Destination     DepartureTime
##  Length:474                999           :  2   12pm - 6pm:205  
##  Class :character          China         :131   6am - 12pm: 46  
##  Mode  :character          Japan         :155   6pm - 9pm :194  
##                            Other         : 22   9pm - 6am : 29  
##                            Southeast Asia:164                   
##                                                                 
##                                                                 
##        SeatClass      Airfare       NoTransport      ModeTransport
##  999        :  4   Min.   :  3.00   1:339       Other       :  2  
##  Business   : 34   1st Qu.: 36.22   2:120       Private Car :106  
##  Economy    :424   Median : 46.83   3: 14       Private Taxi:176  
##  First Class: 12   Mean   : 61.20   4:  1       Public Bus  : 50  
##                    3rd Qu.: 58.00               Public Train: 39  
##                    Max.   :999.00               Subway      :101  
##                                                                   
##    AccessCost       AccessTime        Distance                 Province  
##  Min.   :     0   Min.   :  4.00   Min.   : 23.72    Seoul         :182  
##  1st Qu.:   999   1st Qu.: 30.00   1st Qu.: 46.00    Kyungki-do    :126  
##  Median :  1300   Median : 50.00   Median : 61.00    Jeju          : 89  
##  Mean   :  6785   Mean   : 62.36   Mean   :154.93    Kyungsang-do  : 27  
##  3rd Qu.:  8000   3rd Qu.: 64.00   3rd Qu.:363.00    Incheon       : 22  
##  Max.   :350000   Max.   :390.00   Max.   :455.00    Chungcheong-do: 13  
##                                                     (Other)        : 15  
##  Travel Time, minutes Average Price        a_fare      
##  Min.   : 39          Min.   :   999   Min.   : 14.00  
##  1st Qu.: 56          1st Qu.:   999   1st Qu.: 42.17  
##  Median : 64          Median :   999   Median : 46.83  
##  Mean   :103          Mean   :166373   Mean   : 61.20  
##  3rd Qu.:180          3rd Qu.:362160   3rd Qu.: 52.72  
##  Max.   :311          Max.   :529260   Max.   :999.00  
##

Observations

5.4 Univariate graphical – Quantitative

5.4.1 Histograms

# Code histograms using ggplot() so we can see all quant variables

grid.arrange(
  airport_survey_data %>% 
  ggplot(aes(Age)) +
    geom_histogram(fill = "darkolivegreen3") +
  xlab("Age (in years)")+
  theme_classic(),
  
  airport_survey_data %>% 
  ggplot(aes(Airfare)) +
    geom_histogram(fill = "darkolivegreen3") +
  xlab("Airfare (in 10,000 KRW)")+
  theme_classic(),
  
  airport_survey_data %>% 
  ggplot(aes(AccessCost/1000)) +
    geom_histogram(fill = "darkolivegreen3") +
  xlab("Transportation Cost to the Airport (in 1000 KRW)") +
  theme_classic() +
  xlim(0, 100) +
  ylim(0, 100),
  
  airport_survey_data %>% 
  ggplot(aes(AccessTime)) +
    geom_histogram(fill = "darkolivegreen3") +
  xlab("Transportation Time to the Airport (in minutes)")+
  theme_classic()
)

5.4.2 Observations

5.4.3 Boxplots of quantitative data

# get box plot of all the quanitative variables
par(mfrow = c(2, 2))
boxplot(Age, main = "Age (in years)")
boxplot(Airfare, main = "Airfare (in 10,000 KRW)")
boxplot(AccessCost/1000, main = "Transportation Cost to Airport(1000KRW)")
boxplot(AccessTime, main = "Transportation Time to Airport(in min)")

5.4.4 Observations

5.5 Multivariate non-graphical – Categorical

5.5.1 Cross-Tabs with Airport

5.5.1.1 Airport and Airlines Count

# Frequency count
dat1 %>%
  tabyl(Airport, Airline)

##  Airport Asiana Airlines Foreign Airlines Korean Air Korean LCC
##      GMP              52               93         76         24
##      ICN              54               44         74         57

# Make cross-tabs between two categorical variables using janitor package
# cross-tab of bedrooms and floors
airport_survey_data %>% 
  tabyl(Airline, Airport) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Airport
Airline	GMP	ICN	Total
Asiana Airlines	10.97%	11.39%	22.36%
Foreign Airlines	19.62%	9.28%	28.90%
Korean Air	16.03%	15.61%	31.65%
Korean LCC	5.06%	12.03%	17.09%
Total	51.69%	48.31%	100.00%

5.5.1.2 Airport and Gender Count

# Frequency count
dat1 %>%
  tabyl(Airport, Gender)

##  Airport Female Male Not Mentioned
##      GMP    134  110             1
##      ICN    115  112             2

# cross-tab of bedrooms and waterfront

airport_survey_data %>% 
  tabyl(Gender, Airport) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2)%>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Airport
Gender	GMP	ICN	Total
Female	28.27%	24.26%	52.53%
Male	23.21%	23.63%	46.84%
Not Mentioned	0.21%	0.42%	0.63%
Total	51.69%	48.31%	100.00%

Females prefer GMP Airport
Males have almost equal percent of travelling via GMP and ICN airports

5.5.1.3 Airport and Nationality Count

# Frequency count
dat1 %>%
  tabyl(Airport, Nationality)

##  Airport China Japan Korea Other Southeast Asia
##      GMP    26    39   171     7              2
##      ICN     6     1   197     7             18

5.5.1.4 Airport and Province Residence Count

# Frequency count
dat1 %>%
  tabyl(Airport, Province)

##  Airport  Chungcheong-do  Incheon  Jeju  Jeonra-do  Kangwon-do  Kyungki-do
##      GMP               6       11    69          5           3          59
##      ICN               7       11    20          5           2          67
##   Kyungsang-do  Seoul
##              5     87
##             22     95

# cross-tab of bedrooms and condition
airport_survey_data %>% 
  tabyl(Province, Airport) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Airport
Province	GMP	ICN	Total
Chungcheong-do	1.27%	1.48%	2.74%
Incheon	2.32%	2.32%	4.64%
Jeju	14.56%	4.22%	18.78%
Jeonra-do	1.05%	1.05%	2.11%
Kangwon-do	0.63%	0.42%	1.05%
Kyungki-do	12.45%	14.14%	26.58%
Kyungsang-do	1.05%	4.64%	5.70%
Seoul	18.35%	20.04%	38.40%
Total	51.69%	48.31%	100.00%

ICN Airport is preferred by Seoul Province
GMP Airport is preferred by Jeju Province

5.5.1.5 Airport and Destination Count

# Frequency count
dat1 %>%
  tabyl(Airport, Destination)

##  Airport 999 China Japan Other Southeast Asia
##      GMP   2    87   130     4             22
##      ICN   0    44    25    18            142

# Make cross-tabs between two categorical variables using janitor package
# cross-tab of Airport and Destinationn
airport_survey_data %>% 
  tabyl(Destination, Airport) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2)%>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Airport
Destination	GMP	ICN	Total
999	0.42%	0.00%	0.42%
China	18.35%	9.28%	27.64%
Japan	27.43%	5.27%	32.70%
Other	0.84%	3.80%	4.64%
Southeast Asia	4.64%	29.96%	34.60%
Total	51.69%	48.31%	100.00%

5.5.1.6 Airport and Frequent flight destination

# cross-tab of freq flight destination and Airport
airport_survey_data %>% 
  tabyl(FrequentFlightDestination, Airport) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Airport
FrequentFlightDestination	GMP	ICN	Total
999	2.53%	4.01%	6.54%
China	16.24%	5.27%	21.52%
Europe	1.90%	3.59%	5.49%
Japan	18.99%	4.22%	23.21%
None	1.05%	1.48%	2.53%
North/South America	1.27%	1.27%	2.53%
Other	3.80%	2.53%	6.33%
Southeast Asia	5.91%	25.95%	31.86%
Total	51.69%	48.31%	100.00%

With approximately 19%, Japan is the frequent flight destination from GMP

5.5.1.7 Airport and Departure Time Count

# Frequency count
dat1 %>%
  tabyl(Airport, DepartureTime)

##  Airport 12pm - 6pm 6am - 12pm 6pm - 9pm 9pm - 6am
##      GMP        135         35        74         1
##      ICN         70         11       120        28

5.5.2 Cross-Tabs with Airline

5.5.2.1 Airline and Gender Count

# Frequency count
dat1 %>%
  tabyl(Airline, Gender)

##           Airline Female Male Not Mentioned
##   Asiana Airlines     52   53             1
##  Foreign Airlines     71   66             0
##        Korean Air     77   73             0
##        Korean LCC     49   30             2

# cross-tab of Airline and Gender
airport_survey_data %>% 
  tabyl(Airline, Gender) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Gender
Airline	Female	Male	Not Mentioned	Total
Asiana Airlines	10.97%	11.18%	0.21%	22.36%
Foreign Airlines	14.98%	13.92%	0.00%	28.90%
Korean Air	16.24%	15.40%	0.00%	31.65%
Korean LCC	10.34%	6.33%	0.42%	17.09%
Total	52.53%	46.84%	0.63%	100.00%

5.5.2.2 Airline and Nationality Count

# Frequency count
dat1 %>%
  tabyl(Airline, Nationality)

##           Airline China Japan Korea Other Southeast Asia
##   Asiana Airlines     5     8    84     4              5
##  Foreign Airlines    18    20    90     4              5
##        Korean Air     9    10   123     4              4
##        Korean LCC     0     2    71     2              6

# cross-tab of Airline and Nationality
airport_survey_data %>% 
  tabyl(Airline, Nationality) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Nationality
Airline	China	Japan	Korea	Other	Southeast Asia	Total
Asiana Airlines	1.05%	1.69%	17.72%	0.84%	1.05%	22.36%
Foreign Airlines	3.80%	4.22%	18.99%	0.84%	1.05%	28.90%
Korean Air	1.90%	2.11%	25.95%	0.84%	0.84%	31.65%
Korean LCC	0.00%	0.42%	14.98%	0.42%	1.27%	17.09%
Total	6.75%	8.44%	77.64%	2.95%	4.22%	100.00%

5.5.2.3 Airline and Nationality Count

# Frequency count
dat1 %>%
  tabyl(Airline, Destination)

##           Airline 999 China Japan Other Southeast Asia
##   Asiana Airlines   0    32    47     3             24
##  Foreign Airlines   0    46    53    10             28
##        Korean Air   2    53    48     9             38
##        Korean LCC   0     0     7     0             74

# cross-tab of Airline and Destination
airport_survey_data %>% 
  tabyl(Airline, Destination) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Destination
Airline	999	China	Japan	Other	Southeast Asia	Total
Asiana Airlines	0.00%	6.75%	9.92%	0.63%	5.06%	22.36%
Foreign Airlines	0.00%	9.70%	11.18%	2.11%	5.91%	28.90%
Korean Air	0.42%	11.18%	10.13%	1.90%	8.02%	31.65%
Korean LCC	0.00%	0.00%	1.48%	0.00%	15.61%	17.09%
Total	0.42%	27.64%	32.70%	4.64%	34.60%	100.00%

5.5.2.4 Airline and DepartureTime Count

# Frequency count
dat1 %>%
  tabyl(Airline, DepartureTime)

##           Airline 12pm - 6pm 6am - 12pm 6pm - 9pm 9pm - 6am
##   Asiana Airlines         39          4        55         8
##  Foreign Airlines         60          8        61         8
##        Korean Air         89         15        43         3
##        Korean LCC         17         19        35        10

# cross-tab of Airline and DepartureTime
airport_survey_data %>% 
  tabyl(Airline, DepartureTime) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	DepartureTime
Airline	12pm - 6pm	6am - 12pm	6pm - 9pm	9pm - 6am	Total
Asiana Airlines	8.23%	0.84%	11.60%	1.69%	22.36%
Foreign Airlines	12.66%	1.69%	12.87%	1.69%	28.90%
Korean Air	18.78%	3.16%	9.07%	0.63%	31.65%
Korean LCC	3.59%	4.01%	7.38%	2.11%	17.09%
Total	43.25%	9.70%	40.93%	6.12%	100.00%

5.5.3 Trip Purpose and Destination

# cross-tab of floors and waterfront
airport_survey_data %>% 
  tabyl(TripPurpose, Destination) %>% 
  adorn_totals(where = c("row", "col")) %>%
  adorn_percentages(denominator = "all") %>%
  adorn_pct_formatting(digits = 2) %>%
  adorn_title() %>%
   kable() %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  row_spec(1, background = "#e5f5e0")

	Destination
TripPurpose	999	China	Japan	Other	Southeast Asia	Total
Business	0.00%	7.81%	9.70%	0.84%	3.16%	21.52%
Leisure	0.21%	14.77%	18.78%	3.38%	29.75%	66.88%
Other	0.21%	2.74%	1.69%	0.21%	1.27%	6.12%
Study	0.00%	2.32%	2.53%	0.21%	0.42%	5.49%
Total	0.42%	27.64%	32.70%	4.64%	34.60%	100.00%

5.5.4 Observations

Japan is mostly travelled by the passengers for business purposes
Southeast Asia is favorite for Leisure purpose

5.6 Multivariate graphical code – Categorical

#Function for heatmap
 heat_map <- function (x, y,y_title,x_title){
   x<-enquo(x)
   y<-enquo(y)
   
   
   airport_survey_data %>%
     group_by(!!x,!!y) %>%
     summarise(count = n())  %>%
     ggplot(aes(!!x,!!y)) +
     geom_tile(aes(fill = count))+
     ylab(y_title) + xlab(x_title)+
     theme_classic() +
     scale_fill_continuous(guide = guide_legend(title = "Count")) +
     scale_fill_gradient(low = "white", high = "darkolivegreen3")
     
 }

grid.arrange(
heat_map(Airport, Airline, "Airline", "Airport"),
heat_map(Airport, Destination, "Destination", "Airport"),
heat_map(Airport, Province, "Province", "Airport"),
heat_map(TripPurpose, Destination, "Destination", "TripPurpose"),
nrow=2)

heat_map(Airport, TripPurpose, "TripPurpose", "Airport" )

heat_map(Airline, TripPurpose, "TripPurpose", "Airline" )

heat_map(GroupTravel, Airline, "GroupTravel", "Airport" )

heat_map(GroupTravel, Airport, "GroupTravel", "Airline" )

heat_map(Gender, Airline, "Airline", "Gender" )

heat_map(Gender, Airport, "Airport", "Gender" )

ggplot(dat1, aes(x=Age, y=Airline)) +
  geom_point(size=2, shape=23)

ggplot(dat1, aes(x=Age, y=Airport)) +
  geom_point(size=2, shape=23)

ggplot(dat1, aes(x=Age, y=Gender, color=Airport)) +
  geom_point(size=2, shape=23)

ggplot(dat1, aes(x=Age, y=Gender, color=Airline)) +
  geom_point(size=2, shape=23)

ggplot(dat1, aes(x=Destination, y=Airfare, color=Airline)) +
  geom_point(size=2, shape=23) + ylim(0, 270)

## Warning: Removed 6 rows containing missing values (geom_point).

5.7 Multivariate non-graphical – Quantitative

The standard measure between quantitative variables is correlation

5.7.1 Correlation plots

# Find correlation of quantitative variables
#cor_plot <- dat1 %>%
#  select(-c(FrequentFlightDestination, -MileageAirline))

#cor(cor_plot)

5.7.2 Observations

5.8 Multivariate graphical code – Quantitative

5.8.1 Pairwise Plot

# Correlation plot
dat1 %>%
  select(Airport, Airline, Age, Gender, TripPurpose, TripDuration, Destination, DepartureTime, Province, Airfare, AccessCost, AccessTime) %>%   
  pairs.panels(hist.col="darkolivegreen3")

6 Detailed EDA

6.1 Airlines vs Airfare

airport_airline_airfare_graph <- ggplot(airport_survey_data, aes(x = Airline, y = Airfare, color = Airport)) + 
geom_point(alpha = 0.5) +
   ylim(0, 300) +
  geom_smooth(method=lm, se=FALSE) +
theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0, face = 'bold',color = 'black'),
      plot.subtitle = element_text(face = "italic")) +
labs(x = 'Airlines', y = 'Airfare  (10,000 Korean Won, KRW)', title = "ICN airport has higher airfares") +
  scale_fill_brewer(palette = "blue")

 airport_airline_airfare_graph

airport_airline_bargraph <- airport_survey_data %>%
  group_by(Airport, Airline) %>%
  summarize (Avg_Fare = mean(Airfare)) 

ggplot(data = airport_airline_bargraph, aes(x = Airline, y = Avg_Fare, fill = Airport)) +
  geom_bar(stat="identity", position="dodge") +
  coord_flip() +
    scale_fill_brewer() +
  labs(x = 'Airlines', y = 'Airfare  (10,000 Korean Won, KRW)', title = "ICN airport has higher average airfares") +
  theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0, face = 'bold',color = 'black'),
      plot.subtitle = element_text(face = "italic"))

airport_gender_age_graph <- ggplot(airport_survey_data, aes(x = Gender, y = Age, color = Airport)) + 
geom_point(alpha = 0.5) +
   ylim(0, 85) +
  geom_smooth(method=lm, se=FALSE, color="black") +
theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0, face = 'bold',color = 'black'),
      plot.subtitle = element_text(face = "italic")) +
labs(x = 'Gender', y = 'Age (Years)', title = "Older people prefer GMP airport") +
guides(color = guide_legend(title = '  Airports')) +
  scale_fill_brewer()

 airport_gender_age_graph

airport_gender_age_bargraph <- airport_survey_data %>%
  group_by(Airport, Gender) %>%
  summarize (Avg_age = mean(Age)) 

ggplot(data = airport_gender_age_bargraph, aes(x = Gender, y = Avg_age, fill = Airport)) +
  geom_bar(stat="identity", position="dodge") +
  coord_flip() +
     scale_fill_brewer()  +
  labs(x = 'Gender', y = 'Age (Years)', title = "Among surveyed, average age of people going to GMP is higher") +
  theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0, face = 'bold',color = 'black'),
      plot.subtitle = element_text(face = "italic"))

airport_destination_bargraph <- airport_survey_data %>%
  group_by(Airport, Destination) %>%
  summarize (Count = n())

ggplot(data = airport_destination_bargraph, aes(x = Destination, y = Count, fill = Airport)) +
  geom_bar(stat="identity", position="dodge") +
  coord_flip() +
     scale_fill_brewer()  +
  labs(x = 'Destination', y = 'Count', title = "People prefer GMP airport while travelling to Japan and China", subtitle = "People prefer ICN airport while travelling to South East Asia and other airports") +
  theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0, face = 'bold',color = 'black'),
      plot.subtitle = element_text(face = "italic"))

airport_province_bargraph <- airport_survey_data %>%
  group_by(Airport, Province) %>%
  summarize (Count = n())

ggplot(data = airport_province_bargraph, aes(x = Province, y = Count, fill = Airport)) +
  geom_bar(stat="identity", position="dodge") +
  coord_flip() +
     scale_fill_brewer()  +
  labs(x = 'Province', y = 'Count', title = "More travelled via GMP from Seoul,Kyungsang-do,Kyungki-do,Chungcheong-do", subtitle = "More travelled via ICN from Kangwon-do and Jeju") +
  theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0.6, color = 'black'),
      plot.subtitle = element_text(face = "italic"))

airport_transport_travelTime_graph <- airport_survey_data %>%
  group_by(Airport, ModeTransport) %>%
  summarize (Average_Travel_Time = mean(`Travel Time, minutes`))

ggplot(data = airport_transport_travelTime_graph, aes(x = ModeTransport, y = Average_Travel_Time, fill = Airport)) +
  #geom_bar(stat="identity", position="dodge") +
  geom_bar(stat = "identity", position = position_dodge2(preserve = "single", padding = 0)) +
  coord_flip() +
      scale_fill_brewer() +
  labs(x = 'Mode of Transport', y = 'Average Travel Time (in minutes)', title = "Average Travel Time to ICN is more") +
  theme_classic()+
theme(legend.title = element_text(size=10),
      plot.title = element_text(hjust = 0.6, face = 'bold', color = 'black'),
      plot.subtitle = element_text(face = "italic"))

ggplot(dat1, aes(fill=Airline, y=Airfare, x=Airline)) + 
    geom_bar(position="dodge", stat="summary", fun.y="mean")

6.1.1 Observations

7 Statistical EDA

7.1 Chi-squared test

7.1.1 Airport and Airline

chisq.test(table(airport_survey_data$Airport, airport_survey_data$Airline))

## 
##  Pearson's Chi-squared test
## 
## data:  table(airport_survey_data$Airport, airport_survey_data$Airline)
## X-squared = 30.529, df = 3, p-value = 1.068e-06

7.1.2 Airport and Age

chisq.test(table(airport_survey_data$Airport, airport_survey_data$Age))

## 
##  Pearson's Chi-squared test
## 
## data:  table(airport_survey_data$Airport, airport_survey_data$Age)
## X-squared = 77.248, df = 57, p-value = 0.03835

7.1.3 Airport and Destination

chisq.test(table(airport_survey_data$Airport, airport_survey_data$Destination))

## 
##  Pearson's Chi-squared test
## 
## data:  table(airport_survey_data$Airport, airport_survey_data$Destination)
## X-squared = 183.63, df = 4, p-value < 2.2e-16

7.1.4 Airport and Airfare

chisq.test(table(airport_survey_data$Airport, airport_survey_data$Airfare))

## 
##  Pearson's Chi-squared test
## 
## data:  table(airport_survey_data$Airport, airport_survey_data$Airfare)
## X-squared = 237.84, df = 90, p-value = 2.739e-15

7.2 t-test

t.test(airport_survey_data$Airfare[airport_survey_data$Airport == "ICN"], airport_survey_data$Airfare[airport_survey_data$Airport == "GMP"])

## 
##  Welch Two Sample t-test
## 
## data:  airport_survey_data$Airfare[airport_survey_data$Airport == "ICN"] and airport_survey_data$Airfare[airport_survey_data$Airport == "GMP"]
## t = 1.5608, df = 401.61, p-value = 0.1194
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -4.102883 35.718697
## sample estimates:
## mean of x mean of y 
##  69.37374  53.56583

t.test(airport_survey_data$Age[airport_survey_data$Airport == "ICN"], airport_survey_data$Age[airport_survey_data$Airport == "GMP"])

## 
##  Welch Two Sample t-test
## 
## data:  airport_survey_data$Age[airport_survey_data$Airport == "ICN"] and airport_survey_data$Age[airport_survey_data$Airport == "GMP"]
## t = -3.1916, df = 467.59, p-value = 0.00151
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -6.338893 -1.507751
## sample estimates:
## mean of x mean of y 
##  37.79913  41.72245

t.test(airport_survey_data$`Travel Time, minutes`[airport_survey_data$Airport == "ICN"], airport_survey_data$`Travel Time, minutes`[airport_survey_data$Airport == "GMP"])

## 
##  Welch Two Sample t-test
## 
## data:  airport_survey_data$`Travel Time, minutes`[airport_survey_data$Airport ==  and airport_survey_data$`Travel Time, minutes`[airport_survey_data$Airport ==     "ICN"] and     "GMP"]
## t = 0.82831, df = 454.09, p-value = 0.4079
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -8.255885 20.285900
## sample estimates:
## mean of x mean of y 
##  106.1048  100.0898

7.3 ANOVA test

# Compute the analysis of variance
res.aov <- aov(Airfare ~ Airport + Airline + Gender + Destination, data = airport_survey_data)
# Summary of the analysis
summary(res.aov)

##              Df  Sum Sq Mean Sq F value Pr(>F)   
## Airport       1   29578   29578   2.553 0.1107   
## Airline       3   59951   19984   1.725 0.1610   
## Gender        2    1159     580   0.050 0.9512   
## Destination   4  170079   42520   3.670 0.0059 **
## Residuals   463 5363580   11584                  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

7.3.1 Preparing the Data

summary(dat1)

##    Airport            Airline               Age           Gender         
##  Length:474         Length:474         Min.   :17.00   Length:474        
##  Class :character   Class :character   1st Qu.:29.00   Class :character  
##  Mode  :character   Mode  :character   Median :37.00   Mode  :character  
##                                        Mean   :39.83                     
##                                        3rd Qu.:50.00                     
##                                        Max.   :80.00                     
##                                                                          
##  Nationality        TripPurpose         TripDuration     GroupTravel    
##  Length:474         Length:474         Min.   :  0.00   Min.   :0.0000  
##  Class :character   Class :character   1st Qu.:  4.00   1st Qu.:0.0000  
##  Mode  :character   Mode  :character   Median :  5.00   Median :0.0000  
##                                        Mean   : 24.59   Mean   :0.1624  
##                                        3rd Qu.:  8.00   3rd Qu.:0.0000  
##                                        Max.   :730.00   Max.   :1.0000  
##                                                                         
##  FrequentFlightDestination Destination        DepartureTime     
##  Length:474                Length:474         Length:474        
##  Class :character          Class :character   Class :character  
##  Mode  :character          Mode  :character   Mode  :character  
##                                                                 
##                                                                 
##                                                                 
##                                                                 
##   SeatClass            Airfare        NoTransport    ModeTransport     
##  Length:474         Min.   :  3.00   Min.   :1.000   Length:474        
##  Class :character   1st Qu.: 36.22   1st Qu.:1.000   Class :character  
##  Mode  :character   Median : 46.83   Median :1.000   Mode  :character  
##                     Mean   : 61.20   Mean   :1.319                     
##                     3rd Qu.: 58.00   3rd Qu.:2.000                     
##                     Max.   :999.00   Max.   :4.000                     
##                                                                        
##    AccessCost       AccessTime        Distance                 Province  
##  Min.   :     0   Min.   :  4.00   Min.   : 23.72    Seoul         :182  
##  1st Qu.:   999   1st Qu.: 30.00   1st Qu.: 46.00    Kyungki-do    :126  
##  Median :  1300   Median : 50.00   Median : 61.00    Jeju          : 89  
##  Mean   :  6785   Mean   : 62.36   Mean   :154.93    Kyungsang-do  : 27  
##  3rd Qu.:  8000   3rd Qu.: 64.00   3rd Qu.:363.00    Incheon       : 22  
##  Max.   :350000   Max.   :390.00   Max.   :455.00    Chungcheong-do: 13  
##                                                     (Other)        : 15  
##  Travel Time, minutes Average Price        a_fare      
##  Min.   : 39          Min.   :   999   Min.   : 14.00  
##  1st Qu.: 56          1st Qu.:   999   1st Qu.: 42.17  
##  Median : 64          Median :   999   Median : 46.83  
##  Mean   :103          Mean   :166373   Mean   : 61.20  
##  3rd Qu.:180          3rd Qu.:362160   3rd Qu.: 52.72  
##  Max.   :311          Max.   :529260   Max.   :999.00  
##

summary(airport_survey_data)

##  Airport               Airline         Age                  Gender   
##  GMP:245   Asiana Airlines :106   Min.   :17.00   Female       :249  
##  ICN:229   Foreign Airlines:137   1st Qu.:29.00   Male         :222  
##            Korean Air      :150   Median :37.00   Not Mentioned:  3  
##            Korean LCC      : 81   Mean   :39.83                      
##                                   3rd Qu.:50.00                      
##                                   Max.   :80.00                      
##                                                                      
##          Nationality    TripPurpose   TripDuration     GroupTravel    
##  China         : 32   Business:102   Min.   :  0.00   Min.   :0.0000  
##  Japan         : 40   Leisure :317   1st Qu.:  4.00   1st Qu.:0.0000  
##  Korea         :368   Other   : 29   Median :  5.00   Median :0.0000  
##  Other         : 14   Study   : 26   Mean   : 24.59   Mean   :0.1624  
##  Southeast Asia: 20                  3rd Qu.:  8.00   3rd Qu.:0.0000  
##                                      Max.   :730.00   Max.   :1.0000  
##                                                                       
##  FrequentFlightDestination         Destination     DepartureTime
##  Length:474                999           :  2   12pm - 6pm:205  
##  Class :character          China         :131   6am - 12pm: 46  
##  Mode  :character          Japan         :155   6pm - 9pm :194  
##                            Other         : 22   9pm - 6am : 29  
##                            Southeast Asia:164                   
##                                                                 
##                                                                 
##        SeatClass      Airfare       NoTransport      ModeTransport
##  999        :  4   Min.   :  3.00   1:339       Other       :  2  
##  Business   : 34   1st Qu.: 36.22   2:120       Private Car :106  
##  Economy    :424   Median : 46.83   3: 14       Private Taxi:176  
##  First Class: 12   Mean   : 61.20   4:  1       Public Bus  : 50  
##                    3rd Qu.: 58.00               Public Train: 39  
##                    Max.   :999.00               Subway      :101  
##                                                                   
##    AccessCost       AccessTime        Distance                 Province  
##  Min.   :     0   Min.   :  4.00   Min.   : 23.72    Seoul         :182  
##  1st Qu.:   999   1st Qu.: 30.00   1st Qu.: 46.00    Kyungki-do    :126  
##  Median :  1300   Median : 50.00   Median : 61.00    Jeju          : 89  
##  Mean   :  6785   Mean   : 62.36   Mean   :154.93    Kyungsang-do  : 27  
##  3rd Qu.:  8000   3rd Qu.: 64.00   3rd Qu.:363.00    Incheon       : 22  
##  Max.   :350000   Max.   :390.00   Max.   :455.00    Chungcheong-do: 13  
##                                                     (Other)        : 15  
##  Travel Time, minutes Average Price        a_fare      
##  Min.   : 39          Min.   :   999   Min.   : 14.00  
##  1st Qu.: 56          1st Qu.:   999   1st Qu.: 42.17  
##  Median : 64          Median :   999   Median : 46.83  
##  Mean   :103          Mean   :166373   Mean   : 61.20  
##  3rd Qu.:180          3rd Qu.:362160   3rd Qu.: 52.72  
##  Max.   :311          Max.   :529260   Max.   :999.00  
##

7.3.2 Remove Highly correlated variables

# ProvinceDistance and Access time # Correlated with Province
dat1 <- dat1 %>%
  select(-c(Distance, AccessTime))

# TravelTime, minutes?

Province Distance is highly correlated with Provice, so removed from the dataset
Access Time, time taken by individual from province to airport is also highly correlated with Provice and distance, so removed from the dataset

7.3.3 Remove unrelated variables

# Remove AccessCost - Lot of missing values
# Remove FrequentFlightDestination
# GroupTravel
# TripDuration
# Gender
# AgeGroup
# SeatClass
# NoTransport

dat1 <- dat1 %>%
  select(-c(AccessCost, FrequentFlightDestination, GroupTravel,
            TripDuration, Gender, SeatClass, NoTransport, TripPurpose,
            'Average Price', a_fare, 'Travel Time, minutes', ModeTransport))

## Remove rows with Destination 999
dat1 <- dat1 %>%
  filter(Destination != "999")

## Remove rows with Airfare 999
dat1 <- dat1 %>%
  filter(Airfare != "999")

Remove AccessCost is removed from final dataset - Lot of missing values
Remove FrequentFlightDestination - not used for this study
GroupTravel - not used for this study
TripDuration - not used for this study
Gender - no impact of gender on any of the choice (Airport, Airline)
AgeGroup - no impact of age on any of the choice (Airport, Airline)
SeatClass - Almost all survey participants travelled in economy, so data is biased
NoTransport - Not relevant for this study

8 Save the dataset for use in scikit learn

write.xlsx(dat1, 'export_from_code.xlsx')

## Note: zip::zip() is deprecated, please use zip::zipr() instead

# Write CSV in R
write.csv(dat1, file = "export_from_code.csv",row.names=FALSE)

9 Descriptive Statistics of final dataset

Skim summary statistics
n obs: 466
n variables: 8

Variable type: factor

variable	complete	n	n_unique	top_counts	ordered
Airline	466	466	4	Kor: 148, For: 133, Asi: 105, Kor: 80	FALSE
Airport	466	466	2	GMP: 241, ICN: 225, NA: 0	FALSE
DepartureTime	466	466	4	12p: 203, 6pm: 191, 6am: 43, 9pm: 29	FALSE
Destination	466	466	4	Sou: 160, Jap: 154, Chi: 131, Oth: 21	FALSE
Nationality	466	466	5	Kor: 361, Jap: 39, Chi: 32, Sou: 20	FALSE
Province	466	466	8	Se: 181, Ky: 122, Je: 86, Ky: 27	FALSE

Variable type: numeric

variable	missing	complete	n	mean	sd	p0	p25	p50	p75	p100	hist
Age	0	466	466	39.71	13.59	17	29	37	50	80	▃▇▆▅▃▃▂▁
Airfare	0	466	466	49.24	24.52	3	36.22	46.83	56	260	▃▇▁▁▁▁▁▁

Modeling passenger’s airport and airline choice behavious

Ankita Pathak, Nancy Jain, Sanyukta Ghai, Sourabh Gupta

June 09, 2019

1 Goal of the EDA

2 Prerequisites for analysis

3 Descriptive Statistics

4 Data manipulation

4.1 Categorical Variables - Mutation

4.1.1 Airport

4.1.2 Airline

4.1.3 Gender

4.1.4 Nationality

4.1.5 Trip Purpose

4.1.6 Province Residence

4.1.6.1 Data manipulation in Province raw data file to fill missing values in survey

4.1.6.2 Add province distance and time to the airport_survey_data : Join both raw data

4.1.7 Group Travel

4.1.8 FrequentFlightDestination

4.1.9 Destination

4.1.10 DepartureTime

4.1.11 SeatClass

4.1.12 ModeTransport

4.1.13 Age

4.1.14 Occupation

4.1.15 Income

4.1.16 MileageAirline

4.2 Quantitative Variables - Handle Missing values

4.2.1 Age

4.2.2 Trip Duration

4.2.3 Flying Companion

4.2.4 NoTripsLastYear

4.2.5 Airfare

4.2.6 NoTransport

4.2.7 AccessCost

4.2.8 AccessTime

4.2.8.1 Mileage

4.2.9 Replace remaining NAs by 999 for purpose of EDA

4.3 Adjust Dataset

4.3.1 Removing unnecessary columns

4.3.2 Getting the modified dataset as airport_survey_data

4.3.3 Observations

4.3.4 Export dataset to be used for modeling in Python

5 Basic EDA

5.1 Univariate non-graphical – Categorical

5.1.1 Preferred Airport

5.1.2 Preferred Airline

5.1.3 Gender

5.1.4 Nationality

5.1.5 Trip Purpose

5.1.6 Residence

5.1.7 Destination

5.1.8 Preferred Departure Time

5.1.9 Preferred Seat Class

5.1.10 Preferred Mode of Transport

5.2 Univariate graphical code – Categorical

5.3 Univariate non-graphical – Quantitative

5.3.1 Descriptive Statistics

5.4 Univariate graphical – Quantitative

5.4.1 Histograms

5.4.2 Observations

5.4.3 Boxplots of quantitative data

5.4.4 Observations

5.5 Multivariate non-graphical – Categorical

5.5.1 Cross-Tabs with Airport

5.5.1.1 Airport and Airlines Count

5.5.1.2 Airport and Gender Count

5.5.1.3 Airport and Nationality Count

5.5.1.4 Airport and Province Residence Count

5.5.1.5 Airport and Destination Count

5.5.1.6 Airport and Frequent flight destination

5.5.1.7 Airport and Departure Time Count

5.5.2 Cross-Tabs with Airline

5.5.2.1 Airline and Gender Count

5.5.2.2 Airline and Nationality Count

5.5.2.3 Airline and Nationality Count

5.5.2.4 Airline and DepartureTime Count

5.5.3 Trip Purpose and Destination

5.5.4 Observations

5.6 Multivariate graphical code – Categorical

5.7 Multivariate non-graphical – Quantitative